AI, with added fear
Posted on 20th May 2019
One of the great things about my job is that I learn so much about the modern world and how technology is changing it. The clients we work with are wide-ranging and many are global, not just in the size and scale of their operations, but in their outlook and vision. They way in which they are transforming our world, usually for the better, is, to use a much-abused word, awesome. Let me illustrate with a couple of examples…
Many, many years ago, one of our marketing team was at a dinner for advertising agencies in The Scotsman building on the Bridges in Edinburgh. This was in the days when the Scotsman actually had prestigious offices and sold a decent number of newspapers, but was starting to worry about the impact the internet and tech generally might have on its business.
A lady, sitting opposite him, postulated that in the future a Scotsman (other newspapers are available) reader would have a gel-filled large, thin, envelope/sachet like object which could be rolled up and put in a pocket. When unfurled, it would be possible (she didn’t say how) for the words and pictures of that day’s edition to appear and be read. Folding this package would make it easier to read on the train, plane, or indeed when sitting comfortably anywhere.
My colleague, by no means a technophobe, thought this all a bit unlikely at that time. Yet at the start of this year, I read of a not dissimilar invention; a television that rolls up into a base unit, thus saving space. If you click through on this link you’ll see it, and will note that it’s “coming soon.”
Another use of technology that staggered me was described in The Times at the weekend. We’re teaching cars to be scared. I now know what my colleague felt all those years ago at that Scotsman dinner. How on earth can a lump of metal have feelings? Yet apparently that’s what scientists are doing. More specifically, as The Times reports, “artificial intelligence researchers are … injecting a sense of fear into a self-driving car to help it to make better decisions. A prototype system appears to “feel” something akin to the stab of panic that leads a human driver to hit the brakes when approaching a corner too quickly. The ultimate aim is to build a “visceral machine”, an artificial intelligence imbued with a type of emotional intelligence.”
The thinking behind this is, obviously, to make self-driving cars safer, and it seems to work. When cars are “taught” to drive by AI they learn by having crashes and this teaches them how to avoid accidents. However, when AI with added fear was used to teach a car to learn to drive, it learned more quickly, with 25% fewer crashes than an AI-instructed car with no fear.
There is some amazing tech out there. Much of it we don’t know about yet, either because it’s not been fully developed and/or brought to market or because it’s secret and known only to government agencies. Moreover, as a high-ranking techie at a £Multi-million Scottish company pointed out to me, the brilliant stuff that we read about today has been tested and developed for several years before it makes the pages of, say, The Times of The Scotsman. I wonder what’s coming next? Sadly, for those of a certain age who grew up reading newspapers, I suspect that it’s too late to save the papers, at least in their traditional, newsprint format.
Nikola Kelly, MD, Be-IT
Posted in News
.. Back to Blog